Module 2 of 2 · 40 minutes
Discovery: knowing before you build
By the end of this module you will be able to
- Run discovery that produces decisions rather than decks
- Talk to users in a way that yields truth rather than politeness
- Write and test assumptions cheaply before committing a team
- Kill an idea well, and keep the learning
Work through it
1 interactive for this module, built on the WAJD Teach engine. Nothing moves until you ask it to, and every one has a written version if you would rather read it.
Amara My roadmap is full before any discovery happens. Is that normal?
Nadia It is normal and it is backwards. Delivery answers whether we can build it. Discovery answers whether we should. Skip the second question and you spend three months building what a week of discovery would have killed.
Amara So what does discovery actually consist of?
Nadia Assumptions. Every idea rests on a stack of them. People have this problem. It is painful enough that they act. They would use this solution. They can find it. It is worth what it costs.
Amara And I test them?
Nadia In the right order, which is the part everybody gets wrong. Rank them by how fatal it would be to be wrong, and test the most fatal first.
Amara Why does the order matter so much?
Nadia Because teams naturally test the easy assumptions, confirm them, and feel validated, while the fatal one sits untested at the end. Then it fails in production, expensively, after everything else was proven right.
Amara How cheaply can an assumption be tested?
Nadia Cheaper than you think. A conversation before a prototype. A prototype before a build. A landing page describing the feature before the feature exists, measuring who clicks. Pounds to avoid thousands.
Amara User research. I ask people if they would use things and they always say yes.
Nadia Because people are polite, and politeness ruins research. A yes to would you use this is a kindness, not a data point.
Amara So what do I ask?
Nadia The past, never the future. Walk me through the last time you had this problem. What did you do? What did you try? What did it cost you? Behaviour that already happened is evidence. Predicted behaviour is a favour.
Amara Can I show them my idea?
Nadia Not during the questions. The moment you pitch with enthusiasm, every answer afterwards is contaminated, because they are now helping you rather than informing you. Ask first, show later if at all.
Amara How many people do I need? Surely five is not a sample.
Nadia You are not doing statistics, you are hunting the surprise. The workaround you did not know existed. The step you assumed was painful that is not. The person who already solved it with a spreadsheet. Five or six good conversations usually surface the pattern, and the surprise is worth more than the sample size.
Amara And if discovery says no?
Nadia Then it worked. Discovery that never kills anything is a corridor, not a filter. Killing an idea after a week instead of a quarter is the single largest saving this discipline ever produces.
Amara How do I kill it without burning the person who proposed it?
Nadia Make it about the evidence, never the judgement. Say what was learned, what changed the picture, and what would need to be true for it to come back. And write it down somewhere findable.
Amara Why findable?
Nadia Because the same idea returns every eighteen months with a new sponsor, and the record is the only thing standing between the organisation and paying for the same discovery twice.
The written material
Discovery is deciding what not to build, early and cheaply
Delivery answers: can we build it? Discovery answers: should we? Most organisations skip the second question because the roadmap already promised the feature, which is how teams spend three months building something discovery would have killed in a week.
The unit of discovery is the assumption. Every idea rests on a stack of them: people have this problem, it is painful enough to act on, they would use this solution, they can find it, it is worth what it costs. Write them down and rank them by how fatal it would be to be wrong.
Then test the most fatal one first, as cheaply as possible. A conversation before a prototype, a prototype before a build, a fake door before a feature. The discipline is spending pounds to avoid spending thousands.
Talking to users without leading the witness
People are polite, and politeness ruins research. Ask somebody whether they would use your idea and they will say yes to be kind, and that yes is worth nothing.
Ask about the past, not the future. Not would you use this, but walk me through the last time you had this problem. What did you do? What did you try? What did that cost you? Behaviour that already happened is evidence; predicted behaviour is a favour to your feelings.
Do not pitch during research. The moment you explain your idea with enthusiasm, every answer afterwards is contaminated, because the person is now helping you rather than informing you.
Five or six good conversations usually surface the pattern. You are not seeking statistical significance, you are seeking the surprise: the workaround you did not know about, the step that is not painful at all, the person who solved it with a spreadsheet.
Killing ideas well
Discovery that never kills anything is theatre. If every idea that enters discovery exits into delivery, the process is a corridor, not a filter.
Killing an idea is a success when it costs a week instead of a quarter. Say what was learned, what evidence changed the picture, and what would need to be different for the idea to come back. Then record it somewhere findable, because the same idea returns every eighteen months with a new sponsor, and the record is what stops the organisation paying for the same discovery twice.
And separate the idea from the person. In organisations where a killed idea embarrasses its sponsor, discovery dies quietly, because nobody submits ideas to a process that might humiliate them. The finding is always about the evidence, never about the judgement of whoever proposed it.
Knowledge check
The knowledge check and your certificate need a free account, so that your progress and results can be saved as evidence.
The learning itself stays free and open. You are reading all of it right now without an account.